Popular conceptions of nationhood in old and new European ... ers05.pdf · Popular conceptions of...

Popular conceptions of nationhood in

old and new European member states:

Partial support for the ethnic-civic

framework

Jan Germen Janmaat

Abstract

One of the most influential theories in the study of nationalism has beenthe ethnic-East/civic-West framework developed by Hans Kohn. Usingthe 2002 Eurobarometer survey on national identity and building onearlier survey studies, this article examines whether the Kohn frameworkis valid at the level of popular understandings of nationhood. Itscrutinizes the framework both conceptually �/ do people definenationhood in civic or ethnic terms? �/ and regionally �/ is the Eastindeed more ethnic than the West and the West more civic than the East?It will show that identity markers cluster in a political, a cultural and anethnic dimension. Respondents do not see these dimensions as competingsources of nationhood, however. The article further lends some supportfor the regional component of the framework. Lastly, it argues that it isthe intensity of national identifications rather than their qualitativenature (ethnic-civic) that correlates with xenophobia.

Keywords: Nationalism; civic-ethnic; West-East; national identity; survey data;

attitudes.

The sudden occurrence of (sub-state) nationalist sentiments andviolent ethnic conflicts in Central and Eastern Europe after thecollapse of Communism caught many scholars by surprise. Inaccounting for these phenomena, they rediscovered the ethnic-civicframework developed by Hans Kohn in World War II. Essentially thistheory argues that civic nationalism became the dominant ideology ina few core states in Western Europe and America while ethnic notionsof nationhood prevailed in Eastern Europe and the peripheral areas of

Ethnic and Racial Studies Vol. 29 No. 1 January 2006 pp. 50�/78

# 2006 Taylor & FrancisISSN 0141-9870 print/1466-4356 onlineDOI: 10.1080/01419870500352363

Western Europe. Although few authors embraced this idea whole-heartedly, many of them started using the ethnic-civic terminology andallowed it to influence their writings. From the mid 1990s however agrowing body of literature seriously questioned the validity of theethnic-East/civic-West framework both conceptually and empirically.

Recently a number of studies have appeared that explored the extentto which the ethnic-civic divide is reflected in popular notions ofnationhood. That is, does the population at large define its feeling ofnational belonging in ethnic-civic terms or is the ethnic-civic distinc-tion a purely academic construct, driving academic debates but nothaving any impact on the national affiliations of the common man?Analysing survey data of the 1995 edition of the International SocialSurvey Programme [ISSP], these studies, in brief, showed that people’sconceptions of nationhood indeed clustered in an ascriptive (ethnic)and a voluntarist (civic) dimension. However, they did not findpronounced differences between East and West in the relativeimportance of ethnic and civic criteria.

This study will critically engage and complement these earlierstudies by examining the survey data of the Eurobarometer 2002edition on national identity. It will address a number of questions thatthese studies omitted or only partially investigated. First, it willexplore whether there are underlying dimensions in people’s minds,and if so, to what extent these dimensions coincide with the ethnic-civic dichotomy and relate to those found in the previous studies.

A second question this article addresses is whether civic and ethnicconceptions of nationhood are mutually exclusive or, on the otherhand, reinforcing one another. That is, do people (instinctively) maketradeoffs in a sense that a preference for, say, a civic notion ofnationhood is automatically at the expense of an ethnic one, or arethese notions of nationhood non-competitive? In much of theliterature there is an implicit understanding that ethnic and civicnational identities exclude each other. It is assumed that the twocannot go together because the former is seen as a reflection of liberalinclusive attitudes and the latter as a manifestation of conservatismand xenophobia. But is there tension between the two identities at thelevel of popular understandings of nationhood?

Third, we will examine whether there is a regional difference in thedegree of endorsement of the various identity markers. Stated moredirectly: is the East indeed more ethnic than the West and the Westmore civic than the East? As the Eurobarometer survey includedquestions that covered all aspects of the ethnic-civic distinction, we canexplore its regional component to the fullest extent.

Fourth and last, this study will investigate the relationship betweenidentity markers and xenophobia. Is it true that people whopredominantly support ethnic criteria of nationhood display more

Popular Conceptions of Nationhood 51

negative attitudes towards immigrants than people who endorse civiccriteria, as many experts believe, or is the qualitative nature of nationalidentities irrelevant for opinions on ethnic others?

Analyses of the Eurobarometer data will reveal that the identitymarkers addressed by the ethnic civic framework cluster in threedistinct dimensions: political , cultural and ethnic . The markersclustering in these dimensions are found to be correlating positivelywith one another, which indicates complementary rather thanmutually excluding notions of nationhood. Consequently, we willargue that a three-dimensional model is better suited to representpopular notions of nationhood than a crude ethnic-civic dichotomy ora continuum with ideal types as poles. In addition, and in contrast toprevious studies, this study will show some support for the regionalbreakdown of the ethnic-civic framework. However, doubt is cast onthe stability of this identity pattern. Finally, we will contend that it isthe intensity of national affiliations rather than their qualitative nature(ethnic-civic) that appears to be related to xenofobia and feelings ofcloseness to ethnic others.

First, a brief outline will be given of the ethnic-civic framework andof the criticism it evoked. The second section discusses the results ofthe studies on national identity that used the ISSP survey. Analyses ofthe Eurobarometer data are presented in section four. The articleconcludes by discussing the patterns found and sketching someimplications for existing theories on national identity.

The ethnic-civic dichotomy and its fate in the 1990s and after

Kohn (1944; 1962; 1994) believed that the idea of the nation first arosein countries with a strong bourgeoisie and/or traditions of liberalismand decentralized rule (Great Britain, France, United States, Switzer-land and The Netherlands). This new idea �/ labelled civic nationalismby Kohn �/ inspired millions by propagating the nation as a politicalcommunity of citizens with equal rights and duties. Man was to beliberated from the social bonds: church, class, serfdom, family, thathad kept him ignorant for centuries. Central to the new ideology wasthe notion that every person, irrespective of religious, ethnic or classbackground, could freely join the nation as long as (s)he sworeallegiance to a set of political principles and institutions representingthe nation’s values and objectives. A nationalism of a different kinddeveloped in countries with feudal economies and strong absolutistrule, Kohn argued. Unmitigated by other pressures or interests insociety, this ethnic nationalism commanded an individual’s absolutecommitment to the nation, an attachment overriding all otherloyalties. It rejected the notion of voluntary association and therepresentation of the nation as a modern political community involved

52 Jan Germen Janmaat

in and committed to contemporary social issues. Instead it regardedthe nation as an everlasting natural entity that had slowly evolvedfrom prehistoric times. Membership of the nation was fixed, beinggrounded in descent, native language, religion and customs andfolklore. According to Kohn, the ethnic brand of nationalismprevailed in Germany, Central and Eastern Europe and the peripheryof Western Europe (e.g. Ireland and Spain).

The civic-ethnic distinction has inspired many authors. In an echo ofKohn, Greenfeld and Chirot (1994) identify two types of nations: apolitical nation, which individuals can become part of either by birthor voluntary participation, and an ethnic, collectivistic nation, themembership of which is ascribed by descent (blood). They associatethe first type with Britain and France and the latter with Germany andRussia (see also Hagendoorn and Pepels 2000). In similar vein, writingabout the nation-building projects in the Soviet successor states,Kolstø (2000, p. 2) argues that whereas in the West the nation hastraditionally been understood as a community of citizens held togetherby a common territory, common government authority, a ‘rivalconcept that sees the nation as a cultural entity . . . has deep roots inthe eastern part of Europe, not least in Russia’. Brubaker (1992) usedKohn’s framework to account for differences in citizenship andimmigration policies between France and Germany. Other scholarshave been quick to put the label ethnic on the nationalisms thatfollowed the fall of Communism and the break-up of the SovietUnion, Yugoslavia and Czechoslovakia (e.g. Ignatieff 1994; Snyder1993).1

The ethnic-civic dichotomy has also attracted a great deal ofcriticism. Still relatively mild in his critique is A.D. Smith (1991). Heagrees with Kohn that Western and Eastern models of nationhoodhave different historical roots, but he opposes a crude classificationthat assigns nations to mutually exclusive ethnic and civic categories.Instead he contends that

‘Every nationalism contains civic and ethnic elements in varyingdegrees and different forms. Sometimes civic and territorial elementspredominate; at other times it is the ethnic and vernacularcomponents that are emphasized’ (Smith 1991, p. 13).

Thus, in Smith’s view, the ethnic-civic framework would correspondmore to an ideal type model resembling a continuum with two polesthan to a typology or classification (Kaufmann and Zimmer 2004; seealso Zubrzycki 2001). Every nation would be located somewhere onthis continuum with some occupying a position closer to the civic endand some closer to the ethnic end.


Kymlicka (1999), Nieguth (1999) and Nielsen (1999) have been moredisapproving of the ethnic-civic dichotomy. They have argued that theethnic category, capturing both inclusive and exclusive concepts,should be decomposed into a cultural dimension (language andreligion) which is in principle open to outsiders and an ascriptiveone (kinship, ancestry and race) which is not. Moreover, Nielsen(1999) objects to the term civic nationalism if this is taken to meanliberalism, democracy and state-territorialism and seen as contrastingwith an intolerant antidemocratic ethnic nationalism. Pointing to theLatin American countries and their experience with military dictator-ships, he argues that non-ethnic, territorial nationalisms can be quiteantidemocratic. He therefore concludes: ‘Talk of civic nationalism hadbetter be dropped from our political vocabulary (ibid . p. 127).

Equally critical is Kuzio (2001, 2002). He contends that bothWestern and Eastern nations rest on strong ethnic foundations andthat the former have only become more civic in outlook since the1960s. Drawing on writings of Kaufmann (2000) and A.D. Smith(1998), he advances an evolutionary model that relates the propor-tional mix of civic and ethnic practices in a given state to the age ofthat state and to the consolidation of democratic institutions �/ i.e. theyounger the state and the less opportunity it had to develop a soliddemocracy, the more ethnic it still is (Kuzio 2002).

Schopflin (2000) is particularly outspoken on the ethnic-East/civic-West idea. According to him, ‘the proposition that there is a GoodWestern nationalism (civic, democratic, peace-loving etc) and a BadEastern nationalism (nasty, brutish and anything but short)’ representsa ‘truly lazy’ attitude (ibid . pp. 4,5). Indeed, he may be said to betaking Kuzio’s argument one step further by arguing that from thenineteenth century onwards Western states needed ethnicity to createcredible national communities. These communities, he goes on to say,were necessary for the ever-increasing levels of consent that the statehad to win for its expanding role in society. ‘Without ethnicity’,Schopflin (2000, p. 6) boldly states, ‘it is difficult to secure democracy’.Yet, he does contend that ethnic sentiments are currently stronger inthe East than in the West due to the former region’s particularexperience with Communism:

Communism eliminated all possible civic institutions and codes ofconduct, it turned these societies into civic deserts where the micro-level patterns of behaviour were governed by mistrust and char-acterized by atomization. It was hardly unexpected, therefore, thatethno-nationalism should have acquired the saliency that it did;there was no other identity in the public sphere that could haveplayed this role (ibid. p. 279).


Lastly, Zimmer (2003) made an important contribution to the ethnic-civic debate. He criticizes the ethnic-civic dichotomy for groupinginclusive/exclusive notions and identity markers in one category (eitherethnic or civic). This, he argues, is misleading since it presupposes thatthose who favour a deterministic understanding of nationhood bydefinition use ethno-cultural markers and those who endorse aninclusive vision political markers. In reality, however, the former couldwell rely on political markers and the latter on cultural issues,depending on the issues and political opportunity structure of theday. He therefore proposes to disengage identity markers - ‘symbolicresources’ in his terminology - from the ethnic-civic distinction so thatthe remaining dichotomy reflects exclusively inclusive/voluntarist vs.ascriptive/organic notions of nationhood - ‘boundary mechanisms’ inZimmer’s terms (ibid . p. 178). He further identified four symbolicresources (political values/institutions, culture, history and geography)that political entrepreneurs use to back their inclusive or deterministicvisions of the nation with.

The ISSP survey: No differences between East and West inunderstandings of nationhood

Until the end of the 1990s the ethnic-civic debate had been very mucha theoretical exercise dominated by historians and political scientists.If any empirical data were at all studied these usually involvedstatements by politicians, discussions in the media or policy docu-ments. Little attention was paid to the attitudes and opinions of thecommon man, partly because of a lack of data. The 1995 ISSP survey,however, which focused on national identity and attitudes towardsimmigrants,2 made it possible for the first time to examine to whatextent the ethnic-civic framework is reflected in popular notions ofnationhood. To the knowledge of the author so far four studies haveused this source to explore the popular bases of national identity.Jones and Smith (2001a; 2001b), the authors of two of these studies,have investigated underlying dimensions in people’s minds. They basedtheir analysis on the following question in the survey:

Some people say the following things are important for being [e.g.,truly British, Spanish, Hungarian, etc.]. Others say they are notimportant. How important do you think each of the followingis. . ..?’ 1 (very important), 2 (fairly important), 3 (not very important),or 4 (not important at all).

. To have been born in [respondent’s country].

. To have citizenship in [respondent’s country].

. To have lived in [respondent’s country] for most of one’s life.


. To respect political institutions and laws of [respondent’scountry].

. To feel [British, Spanish, Hungarian, etc.].

. To be able to speak [the dominant language in respondent’scountry].

. To be a [believer in the dominant religion/denomination ofrespondent’s country (e.g., Protestant, Christian, etc.)].

Using a rotated factor analysis they found that the answers to thisquestion clustered in two distinct dimensions, one capturing the itemsborn, citizenship, lived and religion and labelled ‘ascriptive/objectivist’by the authors and another correlating strongly with laws, feel andlanguage which they labelled ‘subjectivist/-voluntarist’. Surprisingly,these dimensions cut right across the ethnic-civic dichotomy astraditionally conceived. The notion of voluntarism, for instance, iscommonly associated with a civic understanding of nationhood butthe voluntarist dimension in Jones and Smith’s study also includeslanguage, an element that is usually grouped in the ethnic category.Likewise, the idea of ascription (i.e. of fixed traits) is attributed to theethnic model of the nation but Jones and Smith’s ascriptive dimensionalso incorporates born, citizenship and lived , political items that arecommonly said to belong to a civic identity. The authors concede,however, that the items may have been understood differently in thevarious national contexts. Thus, whereas place of birth (born) mayhave been associated with the state and its territory in countries likeFrance and the USA which are commonly believed to have strongtraditions of civic nationhood, the same item may have been under-stood as a substitute for descent in countries with reputedly strongerethnic visions of the nation.

In addition, proficiency in the dominant language (language) mighthave been understood as an indicator for integration into the larger(civic) community in migrant nations like Australia and the US, whileit may have been regarded as referring to native language (i.e. anascribed characteristic) in countries that have no tradition ofimmigration. For these reasons, Smith and Jones decided to omitany reference to the ethnic-civic distinction in their characterization ofthe two aforementioned dimensions. Interestingly, their analysis alsorevealed that in most states the ascriptive dimension carried greaterweight than the voluntarist dimension. They therefore conclude: ‘ourfindings suggest an unanticipated homogeneity in the ways thatcitizens around the world think about national identity’ (Jones andSmith 2001a, p. 45), despite ‘distinctive discourses and policies onnational identity, associated with specific religious, social, economicand historical trajectories’ (ibid , p. 58).


A third study using the ISSP survey focused on national pride,national sentiment and xenophobia (Hjerm 2003). The question in thesurvey on national pride was the following:

How proud are you of [country] in each of the following?

1. The way democracy works. �/

2. Its political influence in the world. ½3. [Country’s] economic achievements. ½ political4. Its social security system. ½5. Its fair and equal treatment of ½

all groups in society. �/

6. Its scientific and technologicalachievements. �/

7. Its achievements in sports. ½8. Its achievements in the arts and ½

literature. ½ cultural9. [Country’s] armed forces. ½

10. Its history. �/

Using factor analysis Hjerm distilled two dimensions from the answersto this question, one clustering the first five indicators which heinterpreted as ‘political’, and one grouping the last five indicatorswhich he labelled ‘cultural’. Interestingly, comparing national pridelevels in Eastern and Western Europe, Hjerm found similar levelson the cultural dimension but differing ones on the political dimen-sion, with Western countries exhibiting higher values of politicalnational pride than Eastern states. Although labelling the seconddimension as cultural seems rather far-fetched, clustering as it doesitems as diverse as armed forces, history, and achievements in sport,arts and science, the East-West difference on the more coherentpolitical dimension is noteworthy as it is in line with the traditionalethnic-East/civic-West view.

Unlike the aforementioned studies, the fourth study using the ISSPdata source (Shulman 2002) did not investigate whether the items ofthe survey clustered in one or more dimensions. Focusing on the sameindicators as Jones and Smith and using an additional question on thesharing of traditions,3 Shulman assumed beforehand that the indica-tors born, citizenship, lived, laws and feel reflected a civic under-standing of nationhood and the items language, religion and traditionsa cultural sense of national identity. He took eight countries from thesurvey to represent the West and an equal number to represent theEast. Shulman’s main finding was that there were as many indicatorscontesting the civic-West/ethnic-East argument as supporting it.


In addition, the items that supported the argument showed muchlarger differences within each of the regions than between the tworegions. These findings led Shulman (2002, p. 583) to conclude that:

Imperial and communist rule have not pushed Eastern Europeannationhood in a strongly cultural direction while greatly weakeningcivicness. And whereas most of the West has a long tradition ofdemocracy and relatively strong and stable political institutions,cultural conceptions of nationhood are alive and well.

The studies of both Jones & Smith and Shulman suggest that there areno substantial differences between regions in the strength of civic/voluntaristic or ethnocultural/ascriptive identities. Thus, the East-Westcomponent of the ethnic-civic framework seems to be absent inpopular conceptions of nationhood. Yet, as we have seen, Hjerm’sfindings on national pride do seem to support the East-West divide.This apparent contradiction makes sense if we take a closer look at theitems clustering on Hjerm’s political dimension. Three of these items(the way democracy works, economic achievements, social security)refer to achievements that Western nations can obviously take muchgreater pride in than Eastern states simply because the latter have justbegun developing their democracies and recovering from the post-communist socio-economic crisis.

The Eurobarometer 2002 survey: Non-competitive notions of nationhood

Valuable as the ISSP survey has been in uncovering popularconceptions of nationhood, it nonetheless had two major drawbacks.First, as pointed out by Jones & Smith, a number of items, notablyborn and language, could have been interpreted differently by therespondents, which makes it difficult to assign meaning to underlyingdimensions in the data. Second, the survey did not contain the crucialcriterion of genealogical descent, a fact much deplored by Shulman.Because of this, the ethnic category of the ethnic-civic frameworkcould not be tested to the fullest extent. Moreover, as descent isunequivocally an ascriptive characteristic (Zimmer 2003), it could havegreatly bolstered Jones & Smith’s dimensions if it had been included.The Eurobarometer survey on national identity compensates for theimperfections of the ISSP survey as it does include the item of descent(‘common ancestry’ in the survey).4 Moreover, it also contains the itemof ‘common history and common destiny’, a characteristic that iscommonly associated with an ethnic understanding of nationhood.The survey was held in Great Britain, East and West Germany, Greece,


Italy, Spain, Austria, Hungary, the Czech Republic and Poland andcontained the following question on national identity:

Different things or feelings are crucial to people in their sense ofbelonging to a nation. To what extent do you agree with thefollowing statements?‘‘I feel (NATIONALITY) because I share withmy fellow (NATIONALITY) . . .

1. I do not feel (NATIONALITY)2. A common culture, customs and traditions3. A common language4. Common ancestry5. A common history and a common destiny6. A common political and legal system7. Common rights and duties8. A common system of social security/welfare9. A national economy

10. A national army11. Common borders12. A feeling of national pride13. National independence and sovereignty14. Our national character15. Our national symbols (the flag, the national anthem, etc.)

Respondents were asked to express their level of agreement with eachof the items on a scale with the values of 1 (strongly agree), 2 (tend toagree), 3 (tend to disagree), 4 (strongly disagree).

It must be noted that the question on national identity in theEurobarometer survey was worded quite differently from the one inthe ISSP survey. Whereas the latter asked respondents to evaluate theimportance of certain personal characteristics for making some one atrue [nationality, e.g. Dutchman], the former asked them to state theimportance of certain characteristics buttressing one’s own sense ofnational belonging. The ISSP survey therefore taps much more directlyinto the issue of open or closed membership of the nation andtherefore into Jones & Smith’s voluntarist-ascriptive dichotomy (orZimmer’s boundary mechanisms) than the Eurobarometer. It could beargued that because of its specific wording the Eurobarometer relatesalmost exclusively to properties underpinning national identity(political, societal, cultural, historical, state-symbolic, i.e. the symbolicresources in Zimmer’s terms). In other words, the Eurobarometer dataprimarily cover the qualitative nature of national identities rather thaninclusive or excluding notions of nationhood.


A rotated factor analysis performed on the Eurobarometer datarevealed that the aforementioned items grouped into five dimensions(see Table 1 which presents data of the ten countries groupedtogether5). The first dimension, which is the most powerful in termsof variance explained, could be labelled ‘patriotic’ as it clusters theitems national pride, independence, national character, symbols, and toa lesser extent borders. It is more interesting, however, to examinedimensions two, three and four since the items they cover relate to theethnic-civic framework. As we can see, dimension two neatly groupsitems referring to the political and social system, which makes itplausible to interpret this dimension as political. This result corre-sponds to Hjerm’s findings on national pride and confirms thecontention that there is a distinct political dimension to conceptionsof nationhood. Interestingly, dimensions three (labelled cultural) andfour (labelled ethnic) demonstrate that the items of culture andlanguage should clearly be seen as a separate category distinct fromthe items ancestry and history. This supports the aforementionedcriticism of Kymlicka, Nielsen and Nieguth that the ethnic category as

Table 1. Factor analyses of fourteen national identity items (data from tenEuropean countries; N�/8655)

Item Dimension

1 2 3 4 5

(patriotic) (political) (cultural) (ethnic) (army&borders)

culture 0.278 0.224 0.752 0.312 0.080language 0.208 0.217 0.822 0.256 0.139ancestry 0.257 0.187 0.374 0.758 0.182history 0.297 0.288 0.298 0.762 0.109legal system 0.256 0.762 0.180 0.303 0.126rights and duties 0.267 0.823 0.212 0.166 0.084social security 0.210 0.827 0.151 0.085 0.249economy 0.254 0.677 0.147 0.135 0.472army 0.393 0.363 -0.006 0.239 0.705borders 0.413 0.250 0.326 0.085 0.678national pride 0.776 0.227 0.189 0.202 0.261independence 0.766 0.343 0.176 0.122 0.200national character 0.789 0.266 0.209 0.195 0.164symbols 0.762 0.167 0.188 0.248 0.228

NB: 1. The first item (‘I do not feel [nationality]’) was excluded from the analysis.

2. Items that show loadings of more than 0.5 are underlined.

3. The percentages of variance explained of the five dimensions were 23.1, 22.0, 13.2,

12.0 and 10.9 respectively. The corresponding eigenvalues were 3.24, 3.08, 1.84, 1.68,

and 1.52.


traditionally conceived should be split into a cultural and an ethniccomponent.

It has to be noted though that the pattern of dimensions for thegroup of ten states as a whole does not necessarily correspond to thatof individual states. Not only may the dimensions differ but alsotheir order of importance. Whereas West Germany for instancefollows more or less the general pattern, Hungary presents a stronglycontrasting picture. Not only do the cultural and ethnic itemscluster in one dimension instead of two, also the order of thedimensions is reversed with the ethno-cultural dimension capturingthe highest percentage of variance. Yet, if we group the countries intotwo regions �/ Western Europe and Eastern Europe �/ and carry out theaforementioned factor analysis by region, then both halves of Europedisplay the same pattern of dimensions as the one for all countriestogether.6 Western Europe follows the general pattern precisely, bothin the nature of the dimensions and in their importance, EasternEurope shows the same dimensions but in a slightly different order ofimportance.7 Thus, the dimensions identified in Table 1 have cross-regional validity.

Interestingly, Table 1 also shows that all items correlate positivelywith the five dimensions. This pattern of positive correlations iseven more clearly visible from a table of correlations between pairs ofitems (see Appendix 1 which displays items of both the Eurobarometerand ISSP surveys). The positive correlation between for instanceancestry and rights (0.44) in the Eurobarometer data means that thehigher the level of agreement expressed with common ancestry as anidentity marker so the stronger the support for common rights andduties as a resource underpinning one’s identity. The inference we candraw from this is that people apparently see ancestry and rights, andall the other items in both the Eurobarometer and ISSP surveys, moreas non-competitive complementary concepts than as mutually exclu-sive identity markers. This finding has important consequences fortheories on the nature of the ethnic-civic framework. It shows thatwhen applied to popular notions of nationhood this frameworkcannot be conceived as a dichotomy. Nor can it be viewed as acontinuum with ideal-typical constructs because a continuum alsoimplies competing concepts (the more one moves to one end of thecontinuum the further one moves away from the other end). If we areat all to visualize the conceptual nature of popular understandings ofnationhood, it would be more appropriate to picture these under-standings as a three-dimensional model resembling a cone or apyramid (see Figure 1).

The fact that the models of a dichotomy or a continuum areunsuitable to describe popular notions of nationhood does not meanthat they cannot accurately represent intellectual discourse. Indeed,


contrary to the population at large, many academics, intellectuals andpoliticians do consider ethnic and civic markers to be competingvisions of the nation, the tension between the two visions surfacingwhen citizenship and immigration issues are debated in the media.

Partial support for the regional component of the ethnic-civic framework

Now that we have established that the markers associated with theethnic-civic framework cluster in three dimensions (political, culturaland ethnic), we can examine differences between countries andbetween regions in importance assigned to these dimensions. Table 2presents the mean scores for each country and each region (WesternEurope, Eastern Europe) on the three dimensions.8 The lower thescore, the higher the level of agreement expressed with a particulardimension. Scores higher than 2.5 indicate on average disagreementwith a certain dimension buttressing national identity. The meanscores on the dimensions were calculated by (1) adding up therespondent scores of the items clustering in a dimension, (2) dividingthe resultant figure by the number of these items (producing therespondent dimension scores), and (3) taking the average of therespondent dimension scores. Appendix 2 shows the average scores onthe individual items for the ten countries and the two regions.

The average scores of Table 2 reveal that respondents expressmoderate to high levels of agreement across the board. Not onecountry shows on average disagreement with any of the threedimensions. This high level of consensus means that the differencesbetween the countries and between the regions are not overlyimpressive. Yet, the differences are statistically significant at the 0.01level and do go in the expected direction: East European countriesconsider cultural and ethnic criteria more important markers ofnationhood than the Western countries, the difference between theregions being the largest on the ethnic dimension.9 Vice versa the

Figure 1. Popular understandings of national identity

important

important

political markers

importantcultural markers

ethnic markers

not important


political dimension is deemed less relevant by the Eastern countries.The results are thus clearly in line with the ethnic-East/civic-Westargument.

Nonetheless, it must be noted that the variation between countrieswithin a region is substantial. Thus, Greece, as a West European state,has the lowest score on the ethnic and cultural dimensions of allcountries, which indicates that it attaches more importance to thesedimensions than the four Central European states. Similarly, Poland’saverage score on the political dimension is much lower than that of theother East European states and it is the second lowest score of allcountries. Close scrutiny of the data in fact shows that the respondentsof some countries express high levels of agreement with all threedimensions (e.g. Greece and Poland), whereas the respondents of othercountries are much more reserved across the board (e.g. WestGermany, Italy). This may reflect different attitudes towards socialsurveys and participating in them as a respondent. Respondents insome countries might for instance have felt compelled to providepositive answers, leading them to state high levels of agreement, whilerespondents in other countries may not have felt a similar pressure.

Table 2. Mean scores on dimensions of national identity across ten Europeancountries and two regions

Cultural Ethnic Political

(cultural �/

language)(ancestry �/

history)(legal system �/

rights&duties �/

social security �/

economy)

West Germany 1.69 1.88 1.90Greece 1.28 1.26 1.47Italy 1.64 1.83 1.88Spain 1.57 1.70 1.68Great Britain 1.57 1.72 1.78Austria 1.46 1.81 1.78East Germany 1.54 1.70 1.88Czech Republic 1.48 1.65 2.03Hungary 1.35 1.47 1.92Poland 1.41 1.42 1.56

East 1.44 1.56 1.84West 1.53 1.69 1.74

NB: 1. On each dimension the difference between the mean scores of the two regions is

statistically significant at an 0.01 level.

2. The N of each country score ranges from 766 (Great Britain) to 1013 (Hungary).

Approximately 1000 respondents were interviewed in each country. The lower N of

Great Britain is explained by the many British respondents who indicated that they

did not feel British (item one). We excluded these respondents from the analysis.


In order to control for the propensity to state either high or lowlevels of support for all three dimensions, we subtracted the scores onthe political dimension from those on the ethnic and from those on thecultural dimension (Table 3). The resulting figures indicate thedifference in support between the ethnic and the political dimensionon the one hand, and between the cultural and the political dimensionon the other. The lower the score, the stronger the backing for theethnic or the cultural dimension relative to the political dimension. Aswe can see, the familiar ethnic-East/civic-West pattern emerges evenmore clearly from the data of Table 3. Hungary and the CzechRepublic top the list of both the ethnic-political and cultural-politicalcolumns.10 Moreover, all East European countries are in the top five ofthe ranking list of countries in the ethnic-political column with Greecebeing the only exception. On the cultural-political column the patternis slightly less clear because Poland has a score that is so different fromthe other Eastern states. Aggregated to the regional level, the datashow that there are substantial and statistically significant differencesbetween East and West, with the former attaching much moreimportance to the ethnic and cultural dimensions in relation to thepolitical dimension than the latter.

Yet some important reservations have to be stated. First, EasternEurope is represented by just four countries. In the ISSP data set thevariations between the nine East European countries were at least aslarge as those between the Western states. Thus we might havewitnessed larger differences within Eastern Europe if the Eurobarom-eter survey had included more East European countries. Second, it is

Table 3. Differences of mean scores between the ethnic and the politicaldimension and between the cultural and the political dimension across tenEuropean states and two regions

ethnic minuspolitical

cultural minuspolitical

Hungary �/0.46 Hungary �/0.58Czech Republic �/0.38 Czech Republic �/0.55Greece �/0.21 East Germany �/0.34East Germany �/0.17 Austria �/0.32Poland �/0.14 Italy �/0.25Great Britain �/0.06 West Germany �/0.22Italy �/0.05 Great Britain �/0.21West Germany �/0.02 Greece �/0.19Spain 0.01 Poland �/0.16Austria 0.03 Spain �/0.12

East �/0.29 East �/0.41West �/0.05 West �/0.22


surprising to find West Germany and Austria occupying such lowpositions on the ranking list of countries on the ethnic-politicalcolumn. Austria even has a positive score indicating that itsrespondents show slightly stronger support for political markersthan for ethnic markers. This result is not in line with the ethnic-East/civic-West idea as originally formulated by Kohn as he groupedthe German-speaking countries firmly in the ethnic category. Instead,the fault line in the Eurobarometer data appears to run along theformer Iron Curtain, the former communist states all expressingstronger support for the ethnic dimension vis-a-vis the politicaldimension than the traditional capitalist states of the West (Greecebeing the only exception). Related to this issue is the comparativelylow level of support by the former communist countries for thepolitical dimension. This is not so surprising as it might appear at firstglance if one reconsiders the items of which this dimension iscomposed. As was the case with Hjerm’s political dimension ofnational pride, two of these concern the national economy and thesocial security system. Thus, one could again argue that in view of themodest level of real incomes and social services in former communistcountries in relation to Western countries it is quite logical to find onlylukewarm support for these items in Hungary and the Czech Republic(see Appendix 2).

This conjecture, in turn, raises a number of interesting questions.For instance, have West Germany and Austria always shown such highlevels of support for the political dimension in relation to the ethnicdimension? If so, then Kohn was wrong to assume that an ethnicconception of nationhood prevailed in the German-speaking countriesat the turn of the century (at least as far as mass attitudes on nationalidentity are concerned). If not, then Austrians and West-Germansmust have moved away from a mainly ethnic to a predominantlypolitical view of the nation, a shift that might well be related to thepost-war success of these countries in developing a strong economy, agenerous welfare system and a stable democracy. If it can be provedthat such a process has indeed occurred, then might we not expect thesame identity change to happen in the former communist states as theydevelop their economies and democracies?

Ethnic or civic conceptions of the nation: Do they at all matter?

The reason why the ethnic-civic framework continues to receive somuch scholarly attention is the assumption by many academics andpolicy-makers that civic conceptions of nationhood promote inter-ethnic tolerance and a positive attitude towards immigrants. Bycontrast, ethnic definitions of the nation are seen as contributing toxenophobia and racism. Unfortunately, the Eurobarometer survey did


not contain questions on attitudes towards immigrants. Yet it did askrespondents to state how close they felt towards Jews, Gypsies andvarious other nationalities. This question was worded as follows: ‘Iwould like you to tell me how close you feel to the following groups ofpeople’. Respondents could state their answer on a four-point scaleranging from 1 (very close) to 4 (not at all close). If the aforementionedassumption were true, one would expect the ethnic and, to a lesserextent, the cultural dimensions to display a negative correlation andthe political dimension a positive correlation with feelings of closenesstowards Jews and Gypsies (i.e. the stronger the support for the ethnicand cultural dimensions, the weaker the feelings of closeness; thestronger the support for the political dimension, the stronger thefeelings of closeness).

As we can see in Table 4, this expectation is not at all borne out bythe data. All three dimensions show only very weak correlationswith feelings of closeness, both in the West and in the East (6correlations are not even statistically significant at the 0.05 level).Surprisingly, and totally contrary to the expected relationship, theethnic dimension correlates positively with feelings of closenesstowards Gypsies in both the regions. Another unexpected outcomeis that ethnicity correlates negatively with feelings of closeness towardsJews in the West but positively with these feelings in the East. Thiswould certainly come as a surprise to those who believe that ethnicsentiments and anti-Semitism are closely related phenomena especiallyin Eastern Europe.

As the ISSP survey does include questions on attitudes towardsimmigrants, we can examine whether the pattern of non-correlationsfrom the Eurobarometer also surfaces in the data of that survey. Infact, Hjerm (2003) has already used the ISSP survey to correlateattitudes on immigrants to national pride and national sentiment. Thequestion on immigrants was phrased as follows:

Table 4. Pearson correlations between feelings of closeness towards ethnicothers and dimensions of nationhood in two regions

West East

feeling closeto Jews

feeling closeto Gypsies

feeling closeto Jews

feeling closeto Gypsies

cultural �/0.02 �/0.00 0.05** 0.01ethnic �/0.03* 0.03* 0.03* 0.02political �/0.02 0.03* 0.02 0.06**

*statistically significant at an 0.05 level

**statistically significant at an 0.01 level


There are different opinions about immigrants from other countriesliving in [country]. (By ‘‘immigrants’’ we mean people who came tosettle in [country]). How much do you agree or disagree with each ofthe following statements?

1. Immigrants increase crime rates.2. Immigrants are generally good for [country’s] economy.3. Immigrants take jobs away from people who were born in

[country].4. Immigrants make [country] more open to new ideas and

cultures.

Hjerm constructed an index of what he called ‘xenophobia’ from theanswers to this question by combining the values on the fourstatements (these values ranged from 1-‘agree strongly’ to 5-‘disagreestrongly’ on a five-point scale; the values of statements 1 and 3 werereversed to ensure that they go in the same direction). He found EastEuropean states (Bulgaria, Czech Republic, Hungary, Poland andSlovenia) to exhibit higher xenophobia levels than West Europeanstates (Austria, Britain, Norway, Sweden, The Netherlands). However,correlations with the political and cultural dimensions of nationalpride were weak in both regions, with many countries not showingstatistically significant results (ibid , pp. 420, 421). Although thecorrelations with the political dimension did go in the expecteddirection (negative �/ indicating that the more importance assignedto political items of national pride the lower the level of xenophobia),the correlations with the cultural dimension were negative in two ofthe East European states (not in line with expectation) and positive infour of the five West European states. The first result is in concordancewith our finding that correlations between the three dimensions ofnationhood and feelings of closeness to Jews and Gypsies arenegligible. The last result moreover matches our observation that thecorrelations between ethnicity and feelings of closeness towards Jewsshow contrasting directions in Eastern and Western Europe.

Remarkably, Hjerm did find a relation between xenophobia andnational sentiment , a composite variable he constructed from itemsreferring to feelings of national superiority, unconditional support forone’s country and pride in national achievements.11 Having firstestablished that differences between Eastern and Western countries inthe levels of national sentiment are small, he then shows that nationalsentiment is positively correlated to xenophobia in both regions (i.e.the stronger one’s national sentiment the higher one’s level ofxenophobia). However, the correlations between national sentimentand xenophobia are much stronger in West European states than inEast European countries (ibid. pp. 424, 425). These results suggest twoimportant conclusions: (1) it is not so much the kind of national


identity (ethnic, cultural, political) that matters for perceptions onimmigrants and foreigners, but the strength of national identities andsentiments; (2) Given its stronger connection to xenophobia, nationalsentiment is a more dangerous phenomenon in Western Europe thanin the Eastern half of this continent.

Surprisingly, neither Hjerm nor the other authors who used theISSP data set correlated attitudes on immigrants to conceptions ofnationhood. Combining several databases and making use of multi-level analysis, Jones and Smith (2001b) did relate the voluntarist andascriptive dimensions of national identity to various demographiccharacteristics and to macro-social properties such as a state’s degreeof globalization, its degree of post-industrialism, its degree of internalcultural differentiation and its militarism. They found that all otherthings being equal the higher a country’s degree of post-industrialism,the stronger the support of its population for the voluntaristicdimension. However, they also found strong individual effects, withimmigrants, the higher educated and the well-to-do placing moreimportance on the voluntaristic type than the native-born, the lowereducated and people with modest incomes.

To rectify the aforementioned omission, we used the ISSP databaseto correlate Jones & Smith’s voluntarist and ascriptive dimensions tothe composite construct of xenophobia.12 Obviously, the expectation isthat the voluntarist dimension and xenophobia are negatively corre-lated �/ i.e. respondents who exhibit high scores on this dimensionshould have lower levels of xenophobia than those with low scores.Conversely, those who express strong support for the ascriptivedimension should have higher levels of xenophobia than thoseexpressing more moderate support. The data of Table 5 produce yetanother surprise. The positive correlations between the ascriptivedimension and xenophobia in both Eastern and Western Europe arestill in accordance with the expectation, although it is perhapsremarkable to find Western Europe exhibiting the stronger associa-tion.13 However, what is truly astonishing is that the voluntarist

Table 5. Pearson correlations between xenophobia anddimensions of nation-hood in two regions

West East

xenophobia xenophobia

voluntaristic 0.19** 0.18**ascriptive 0.33** 0.21**

**statistically significant at an 0.01 level


dimension is also positively related to xenophobia in both regions,albeit to a slightly weaker extent than the ascriptive dimension. Thismeans that the stronger a respondent’s support for voluntaristic,inclusive notions of nationhood, the higher that person’s level ofxenophobia. Surely this result must come as an unpleasant surprise tothose who think that strong inclusive national identities are conduciveto more positive opinions on immigrants and foreigners. Takentogether these results provide further evidence for the notion thatattitudes to immigrants and (in)tolerance of other cultures are not somuch dependent on the qualitative nature of national identities butmore on the intensity of these identities.

Discussion and conclusions

This study has examined to what extent the ethnic-civic frameworksurfaces in popular conceptions of nationhood. Using the 2002Eurobarometer survey on national identity it has produced a numberof interesting results that complement and partially support thefindings of studies that utilized the 1995 ISSP survey on nationalidentity.

First of all, we found that the identity markers covered by theethnic-civic framework clustered in three dimensions: ethnic, culturaland political. The identification of a separate cultural factor, distinctfrom the ethnic dimension, is an important new finding. The studiesbased on the ISSP data set could not explore the degree ofinterrelatedness of cultural and ethnic factors because the ISSP surveylacked questions on ethnic markers. As noted before, the existence of aseparate cultural dimension confirms the contention of some critics ofthe ethnic-civic framework that cultural markers of nationhood shouldnot be grouped in the ethnic category because of their inclusive nature.Another point of interest concerns the boundaries between identitymarkers. Whereas our findings conform to the commonly held opinionthat there are distinct political, ethnic and cultural dimensions tonational identifications, Jones & Smith’s voluntarist and ascriptiveconstructs based on the ISSP data cut right across these traditionallyconceived dimensions, the former clustering respect for law, learn thedominant language and self identification and the latter combiningborn, citizenship, length of residence and dominant religion . Thesecontrasting results seem to support Zimmer’s contention that theethnic-civic framework should be disaggregated into two differentmodes of interpreting national identity: one relating to boundarymechanisms and another to symbolic resources. Jones & Smith’sdimensions seem to confirm the existence of boundary mechanisms,while the present study taps into symbolic resources. Zubrzycki’s(2001) study on the debate about the meaning of Polishness in the


media prior to the adoption of the Polish Constitution in April 1997lends further support for Zimmer’s theory. The liberal intelligentsiaengaging in this debate turned to events in the past (i.e. a symbolicresource) to support their argument that Poland had a long traditionof inclusive civic nationhood (i.e. a boundary mechanism).

In fact, our study of the Eurobarometer data revealed that thereare also resources other than political, cultural and ethnic that shapednational allegiances. Four of these clustered in a dimension thatwe labelled patriotism (which proved to be the strongest predictorof national identity) and two remaining items �/ army and commonborders �/ in a construct that could be interpreted as state terri-tory. Thus, the markers of national identity are many and varied, andcan differ from country to country in kind, importance and inter-relatedness.

What they all have in common, however, is their non-competitivenature: not a single pair of markers shows a negative correlation ineither the ISSP or the Eurobarometer database. This is a cardinalfinding. As noted before, many authors assume ethnic and politicalvisions of the nation or voluntarist and ascriptive boundary mechan-isms to be mutually excluding concepts. Remember that Zimmer(2003) argued that the latter could be conceived as a dichotomy. Yet,valid as this interpretation might be for debates among academics,intellectuals and politicians, it evidently does not apply to popularnotions of nationhood. Thus, there may be a distinct differencebetween the elite and the masses in the way they interpret theirnational attachments. The author of the aforementioned study on thePolish constitutional debate also seems to arrive at this conclusionwhen she notes that voter turn-out at the referendum on theConstitution was very low, ‘which seems to indicate that the debatesthat fascinated the elites did not resonate as strongly below’(Zubrzycki 2001, p. 653). Non-competitive notions of nationhoodare in fact not strictly confined to the common man. Zimmer (2003, p.188) himself observes that, in response to the perceived threat fromGermany in the late 1930s, Swiss liberals and conservatives unitedbehind the notion of a distinct Swiss cultural Wesensgemeinschaft ,which effectively ‘blurred the boundary between voluntarist andorganic conceptions of nationhood’. This suggests that the degree inwhich political entrepreneurs disseminate these conceptions as con-trasting visions of nationhood depends very much on the widerpolitical and social circumstances and on the opportunities thesecircumstances offer them. To come back to the issue of elite-massdifferences, it must be noted that these differences extend to otherareas of public opinion as well. Campbell et al . (1960), for instance,observed that the left-right dimension in politics does not by farinfluence the American electorate to the same extent as it does


academics, intellectuals and political professionals. This led them toconclude that ‘the closer the individual stands to the sophisticatedobserver in education and political involvement, the more likely it isthat the observer’s analytical constructs will bear fruit’ (ibid. p. 214). Itis doubtful whether scholars of nationalism realize that their modelsare likely to suffer from the same deficiency.

The results of our analyses also revealed quite strong support for theEast-West component of the ethnic-civic framework. On average EastEuropean respondents accorded more weight to the cultural and,especially, the ethnic dimension and less weight to the politicaldimension than the respondents from Western Europe. This findingappears to somewhat contradict those of Shulman and Jones & Smith.They found either small or no differences at all between the regions,and on some markers the differences were in the opposite directionfrom that expected �/ i.e. the East assigning more importance to some‘civic’ markers than the West (see Shulman 2002, pp. 567, 569).Remember, however, that the ISSP data on which these authors basedtheir analyses did not contain questions on ethnic markers and that thedifference between East and West was most pronounced on the ethnicdimension in the Eurobarometer data. In addition, it must be notedthat Shulman’s civic and cultural categories did not coincide withJones and Smith’s voluntarist and ascriptive dimensions. Thus, someitems that Shulman considered civic were interpreted as ascriptive byJones and Smith. Finally, taking a close look at language and politicalinstitutions, identity markers that were included in both data sets, itcan be noted that the ISSP and the Eurobarometer showed a similarpattern: Eastern Europe places greater value on language but lessvalue on political institutions than does Western Europe (seeAppendix 2 and Shulman 2002, pp. 568, 571). Thus the data of thetwo surveys are consistent.

Obviously, having established the existence of an ethnic-East/political-West pattern in a given year says nothing about its stability.The question to explore is thus to what extent this pattern of identitiesis inert, as primordialists would predict, or subject to change, asinstrumentalists would argue. As the 2001 Eurobarometer was a cross-sectional survey carried out at a single point in time, it does not give usdirect indicators of changes in understandings of nationhood. How-ever, indirectly the data suggest that these understandings are dynamicrather than inert. We noted before that compared to the othercountries in the survey respondents in Austria and West Germanyexpressed strong support for the political dimension relative to theethnic dimension. This is surprising as West Germany and Austriahave traditionally been seen as countries where ethnic conceptions ofnationhood prevail. We postulated that ethnic conceptions mightindeed have been dominant in Austria and West Germany but that the


success of their post-war democracies and socio-economic systemsmay have caused a change in the relative importance of identitymarkers. The post-war political and economic achievements are likelyto have increased trust in and therefore identification with public andpolitical institutions among broad sections of the population. Wefurther suggested that the same change might happen in EastEuropean countries, provided they succeed in developing stabledemocracies and prosperous economies. If such a change happenedin the East, differences between East and West in the qualitative natureof national identifications might disappear altogether.

A change denoting an increased importance of political markerswould strongly support the argument developed by Kuzio that statesbecome more civic in outlook as they grow older and develop stabledemocracies. On close inspection, one could argue that his argument isin fact quite compatible with Kohn’s ethnic-civic model, althoughKuzio would probably strongly deny this himself. Kuzio’s model doesnot rule out the possibility that at some stage in the past ethnicsentiments were stronger among Eastern than Western nations. To theknowledge of the author Kohn never asked himself what wouldhappen to national identifications in a given nation after that nationhad come to define itself ethnically. As Kohn was a historian whoargued that civic nationalism was a new ideology that arose from themiddle classes at the end of the eighteenth century, it is hard toimagine his contending that ethnic identities once established would becast in concrete and thus be resistant to change. If Kohn were in aposition to agree with the argument that ethnic national identities cangradually adopt more civic/political features within the framework ofan independent democratic state, then the opinions of both scholarswould not differ.

This study has to end with a sobering note. Despite the ongoingacademic debate on the ethnic-civic framework and the widely sharedconcern that ethnic nationalism fuels xenophobia and racism, we havenot found the type of national affiliations to really matter for attitudeson immigrants and feelings of closeness to Jews and Gypsies.Remarkably, this applied to both symbolic resources (ethnic, culturaland political markers) and boundary mechanisms (voluntary vsascriptive notions of nationhood). In other words, for their opinionson immigrants and other nationalities, it was irrelevant whetherrespondents cast their national identities in ethnic or political terms,in an inclusive or excluding manner. What did matter was the level ofimportance assigned to the identity markers, i.e. the intensity ofnational identifications. The higher this level, the more negative theperceptions on immigrants. Thus it is not the kind of national identitybut its strength that makes the difference. In line with this argument, itis imaginable that people with more tolerant views on immigrants have


more mixed feelings about national identity in general, predisposingthem to give only half-hearted support to all markers associated withnationhood. In other words, national identity and nationalism may beconsidered improper and unfortunate phenomena altogether by thisgroup of people.

One final reservation has to be stated that might provide somesolace to those convinced of the connection between ethnic/ascribedidentifications and xenophobia. Respondents of both surveys weremost likely not aware of the consequences of each identity marker formembership of the nation when they filled out the questionnaire. Hadthey been informed beforehand of the implications of each item foraccessibility to the nation and of the classifications and judgementsthat invisible academics analysing their responses would make after-wards, they might have shown a different pattern of responses.

Acknowledgements

The author is grateful to the two anonymous ERS referees, and also toStephen Shulman and Andy Green for their useful comments onearlier drafts of this article.

Notes

1. For an extensive overview of scholars that were influenced by the ethnic-civic idea and

of scholars critical of the distinction, see Kuzio (2002) and Shulman (2002).

2. These countries are Ireland, New Zealand, Slovak Republic, The Netherlands, USA,

Canada, Austria, Norway, Australia, Great Britain, Poland, Italy, Latvia, Japan, Sweden,

Russia, Slovenia, Hungary, Czech Republic, Philippines, Spain, Germany (West and East

separately) and Bulgaria.

3. This question was worded as follows: ‘Now we would like to ask you a few questions

about minorities in [respondent’s country]. How much do you agree or disagree with the

following statement: it is impossible for people who do not share the customs and traditions

[of respondent’s country] to become fully [e.g. British, German, Hungarian, etc]?’ The

question had a 5-point agree/disagree scale (agree strongly, agree, neither agree nor disagree,

disagree, disagree strongly ).

4. Between 27 April 2002 and 10 June 2002, the European Opinion Research Group

carried out wave 57.2 of the standard Eurobarometer, on request of the European

Commission, Directorate-General Press and Communication, Opinion Polls. The data of

this survey can be ordered at the Zentralarchiv fuer empirische Sozialforschung of the

Universitaet zu Koeln , Germany.

5. As they constitute separate categories in the survey I hold East and West Germany to

be separate countries.

6. I considered Great Britain, West Germany, Spain, Italy, Austria and Greece to

represent Western Europe, and East Germany, Poland, the Czech Republic and Hungary to

belong to Eastern Europe. Some observers would include Greece in the East European

group, pointing to the long period of Ottoman imperial rule and the lack of civic traditions

and institutions in the country. We chose to include it in Western Europe because of its post-

war (interrupted) development as a democratic state with a market economy, its early


membership of NATO and its inclusion in the EU in the 1981. We concede, however, that this

issue is open to debate.

7. The output of the factor analyses of Germany, Hungary and Western and Eastern

Europe can be obtained from the author.

8. In calculating the scores for Western and Eastern Europe I followed the method

employed by Shulman (2002): all the respondents were first pooled for each region and then

their scores were averaged. The advantage of this approach is that it enables statistical

significance tests of differences between the regions. The disadvantage of some countries

being overrepresented and others underrepresented because of differing sample sizes, which

was a major concern of Shulman, does not apply for the Eurobarometer data because all

countries have approximately equal sample sizes.

9. To assess the statistical significance of differences in the regions’ mean scores an

independent samples t-test was performed on the data.

10. This particular result actually matches the ISSP data set because Hungary and the

Czech Republic headed the ranking list of countries on importance assigned to language in

that data set as well (remember that language is one of the items constituting the cultural

dimension). Moreover, in the same data set respondents in both countries expressed weak

support for the item ‘respect political institutions and law’ (Shulman 2002). This is also

consistent with their score on the political items in the Eurobarometer survey.

11. The question covering national sentiment was phrased as follows: ‘‘How much do you

agree or disagree with the following statements?’’ (‘agree strongly’ �/ ‘disagree strongly’ on 5-

point scale)

�/ I would rather be a citizen of [country] than of any other country in the world.

�/ The world would be a better place if people from other countries were more like those of

[country]

�/ Generally speaking, [country] is a better country than most other countries

�/ People should support their country even if their country is in the wrong

�/ When my country does well in international sports, it makes me proud to be [nationality]

12. We calculated the voluntarist dimension, the ascriptive dimension and the index of

xenophobia in the same way as the dimensions of nationhood in the Eurobarometer data set,

i.e. we added up the respondent scores on the items clustering in a dimension and divided the

resulting figure by the number of items pooled in that dimension.

13. Western Europe includes West Germany, Great Britain, Austria, Italy, Ireland, The

Netherlands, Norway, Sweden and Spain. Eastern Europe is represented by East Germany,

Hungary, Czech Republic, Slovenia, Poland, Bulgaria, Russia, Latvia and Slovak Republic.

References

BRUBAKER, ROGERS 1992 Citizenship and Nationhood in France and Germany,

Cambridge, MA: Harvard University Press

CAMPBELL, ANGUS, et al. 1960 The American Voter, Chicago: The University of

Chicago Press

GREENFELD, LIAH and CHIROT, DANIEL 1994 ‘Nationalism and aggression’, Theory

and society, vol. 23, no. 1, pp. 79�/103

HAGENDOORN, LOUK and PEPELS, JOSE 2000 ‘European nations and nationalism: an

introductory analysis’, in Louk Hagendoorn, et al. (eds), European Nations and Nationalism:

Theoretical and Historical Perspectives , Aldershot: Ashgate, pp. 1�/35

HJERM, MIKAEL 2003 ‘National sentiments in Eastern and Western Europe’, National-

ities Papers , vol. 31, no. 4, pp. 413�/29


IGNATIEFF, MICHAEL 1994 Blood and Belonging: Journeys into the New Nationalism ,

London: Vintage

JONES, FRANK L. and SMITH, PHILIP 2001a ‘Diversity and commonality in national

identities: an exploratory analysis of cross-national patterns’, Journal of Sociology (published

by the Australian Sociological Association), vol. 37, no. 1, pp. 45�/63

*/*/*/*/ 2001b ‘Individual and societal bases of national identity: a comparative multi-level

analysis’, European Sociological Review, vol. 17, no. 2, pp. 103�/118

KAUFMANN, ERIC 2000 ‘Ethnic or civic nation? Theorizing the American case’, Canadian

Review of Studies in Nationalism , vol. 27, no. 1�/2, pp. 133�/55

KAUFFMAN, ERIC and ZIMMER, OLIVER 2004 ‘‘‘Dominant ethnicity’’ and the

‘‘ethnic-civic’’ dichotomy in the work of A.D. Smith’, Nations and Nationalism , vol. 10,

no. 1&2, pp. 63�/78

KOHN, HANS 1944 The Idea of Nationalism: A Study in Its Origins and Background , New

York: Macmillan

*/*/*/*/ 1962 The Age of Nationalism: The First Era of Global History, Westport, Ct:

Greenwood Press

*/*/*/*/ 1994 ‘Western and eastern nationalism’, in John Hutchinson and Anthony D. Smith

(eds), Nationalism , Oxford: Oxford University Press, pp. 162�/65

KOLSTØ, PAL 2000 Political Construction Sites: Nation-Building in Russia and the Post-

Soviet States, Boulder, Colorado: Westview press

KUZIO, TARAS 2001 ‘‘‘Nationalising states’’ or nation-building? A critical review of the

theoretical literature and empirical evidence’, Nations and Nationalism , vol. 7, no. 2, pp.

135�/54

*/*/*/*/ 2002 ‘The myth of the civic state: a critical survey of Hans Kohn’s framework for

understanding nationalism’, Ethnic and Racial studies, vol. 25, no. 1, pp. 20�/39

KYMLICKA, WILL 1999 ‘Misunderstanding nationalism’, in Ronald Beiner (ed.),

Theorizing Nationalism , Albany: State University of New York Press, pp. 131�/40

NIEGUTH, TIM 1999 ‘Beyond dichotomy: Concepts of the nation and the distribution of

membership’, Nations and Nationalism , vol. 5, no. 2, pp. 155�/73

NIELSEN, KAI 1999 ‘Cultural nationalism, neither ethnic nor civic’, in Ronald Beiner (ed.),

Theorizing Nationalism , Albany: State University of New York Press, pp. 119�/30

SCHOPFLIN, GEORGE 2000 Nations, Identity, Power, London: Hurst & Company

SHULMAN, STEPHEN 2002 ‘Challenging the civic/ethnic and West/East dichotomies in

the study of nationalism’, Comparative Political Studies, vol. 35, no. 5, pp. 554�/85

SMITH, ANTHONY D. 1991 National Identity, London: Penguin

*/*/*/*/ 1998 Nationalism and Modernism , London: Routledge

SNYDER, JACK 1993 ‘Nationalism and the crises of the post-Soviet state’, in Michael E.

Brown (ed.), Ethnic Conflict and International Security, Princeton, NJ: Princeton University

Press, pp. 79�/101

ZIMMER, OLIVER 2003 ‘Boundary mechanisms and symbolic resources: towards a

process-oriented approach to national identity’, Nations and Nationalism , vol. 9, no. 2, pp.

173�/193

ZUBRZYCKI, GENEVIEVE 2001 ‘‘We, the Polish Nation’: Ethnic and civic visions of

nationhood in Post-Communist constitutional debates’, Theory and Society, vol. 30, pp.

629�/68

JAN GERMEN JANMAAT is a Post-Doctoral Research Fellow at theSchool of Lifelong Education and International Development, In-stitute of Education, London.ADDRESS: School of LEID, Institute of Education, 20 Bedford Way,London WC1H 0AL, UK. Email: B/[email protected]�/


Appendix 1. Pearson correlations between national identity indicators

Eurobarometer survey

cult lang anc hist leg r&d ss econ arm bor np ind nc sym

culture - .66 .62 .60 .46 .45 .40 .42 .36 .44 .47 .46 .49 .47language - .62 .57 .46 .44 .39 .41 .36 .44 .44 .42 .46 .46ancestry - .72 .48 .44 .40 .44 .43 .45 .51 .47 .51 .50history - .57 .50 .43 .46 .43 .48 .52 .50 .53 .51legal system - .74 .67 .65 .54 .50 .49 .52 .50 .46rights and duties - .73 .65 .51 .50 .48 .52 .49 .44social security - .74 .55 .50 .45 .52 .47 .41economy - .65 .57 .51 .57 .53 .48army - .62 .59 .59 .57 .58borders - .62 .60 .58 .56national pride - .73 .71 .70independence - .72 .66national character - .70symbols -

76

Jan

Germ

enJa

nm

aa

t

Appendix 1 (Continued )

ISSP survey

born citizenship lived law feel language religion traditions

born - .50 .57 .10 .26 .26 .41 .27citizenship - .44 .28 .33 .36 .28 .18lived - .16 .36 .37 .38 .25Law - .28 .28 .16 .10feel - .30 .22 .16language - .21 .22religion - .19traditions -

NB: 1. The N for each correlation ranged from 9210 to 9630 in the Eurobarometer survey and from 28107 to 29870 in the ISSP survey.

2. All the correlations in the two tables are statistically significant at an 0.01 level.

3. ISSP data: we added the separate question on traditions (see note 3) to the other seven items on national identity.

Po

pu

lar

Co

ncep

tion

so

fN

atio

nh

oo

d7

7

Appendix 2. Mean scores on items of national identity across ten European countries and two regions

culture language ancestry history legalsystem

rights &duties

socialsecurity

economy

West Germany 1.79 1.59 1.88 1.89 1.89 1.82 1.92 2.01Greece 1.30 1.26 1.25 1.26 1.43 1.43 1.56 1.51Italy 1.71 1.57 1.75 1.92 1.95 1.78 1.91 1.89Spain 1.61 1.53 1.69 1.71 1.74 1.63 1.67 1.73Great Britain 1.67 1.46 1.71 1.74 1.80 1.73 1.82 1.82Austria 1.51 1.41 1.88 1.75 1.78 1.69 1.74 1.92East Germany 1.61 1.47 1.70 1.71 1.92 1.83 1.88 1.95Czech Republic 1.54 1.43 1.62 1.68 2.05 1.95 2.10 2.05Hungary 1.39 1.31 1.48 1.46 1.84 1.84 2.08 1.96Poland 1.46 1.35 1.42 1.42 1.53 1.56 1.60 1.57

East 1.50 1.39 1.55 1.57 1.83 1.80 1.92 1.88West 1.60 1.47 1.69 1.71 1.76 1.68 1.77 1.81

NB: The N of each country score ranges from 793 to 1016.

78

Jan

Germ

enJa

nm

aa

t

Popular conceptions of nationhood in old and new European ... ers05.pdf · Popular conceptions of...

Documents

Transcript of Popular conceptions of nationhood in old and new European ... ers05.pdf · Popular conceptions of...